Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: study case?

Re: study case?

From: Tuning <c_armanet_at_hotmail.com>
Date: Fri, 30 Aug 2002 17:14:28 +0200
Message-ID: <ako22i$42k$1@wanadoo.fr>


Hi again,

Another question about this parameter, but with Oracle Enterprise Edition 8.1.7, HP-UX :

the change of db_file_multiblock_read_count from 8K to 16K (db_block_size=8K)
doesn't solve the problem of the bad explain plan with CBO; I then thought it could be because
of the HP-UX maximum limit of 64K for I/O operations. Am I right for this point?
Is there a mean to overpass this HP-UX limitation, for this parameter works as well on HP-UX as on Win 2K?

Thanks a lot... And best regards,

C.A.

"Tuning" <c_armanet_at_hotmail.com> a écrit dans le message news: aknu9q$kqi$1_at_wanadoo.fr...
> Hello,
>
> Oracle Enterprise Edition 8.1.7, Win 2K.
>
> I've had a query which was executed with a better explain plan with RBO
than
> CBO (and much faster!)
> with stats on date, with compute option of course!; and I made some
> optimization attempts on init.ora parameters.
> Finaly, CBO worked well with parameter db_multiblock_read_count from 8K to
> 16K.
>
> Then, just for testing, I changed this parameter again, and leave it to
8K.
> So, I was quite surprised when I saw the same explain plan (and fastness)
as
> with 16K!!
>
> Any idea?
>
> Thanks.
>
> Regards,
>
> C.A.
>
>
Received on Fri Aug 30 2002 - 10:14:28 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US